Methodological challenges in psychedelic drug trials: Efficacy and safety of psilocybin in treatment-resistant major depression (EPIsoDE) - Rationale and study design

CONCLUSION

Here we present the rationale and study design of our phase IIb trial on the efficacy and safety of psilocybin in TRD. The study addresses several crucial methodological questions in psychedelic drug trials. Specifically, these are the risk of unblinding and nocebo effects, prevalent "expectation bias", and the choice of the most adequate comparator dose. In clinical trials of psychotropic drugs, the experimental drug is usually compared to a placebo and/or an active comparator with proven efficacy in that same indication. While there has been some debate over the last decade whether the effect size in antidepressant clinical trials against placebo might be exaggerated due to unblinding through the perception of side effects in the verum arm, it is general consensus that classical antidepressant drugs have a superior albeit modest efficacy compared to placebo, and that their efficacy cannot be attributed to the perception of side effects alone. However, the treatment effect in randomized controlled trials (RCTs) is clearly modulated by patients' and therapists' expectations. Response rates are larger and dropout rates smaller in RCTs without a placebo control, and the higher the probability for receiving placebo in a particular trial the smaller the response rate and the higher the dropout rate.conclude from their findings that the probability of receiving placebo should be considered when interpreting and synthesizing results from RCTs in major depression. The effect of this "expectation bias" by both clinicians and patients on outcomes in clinical trials cannot be underestimated; it influences the magnitude of symptom reduction with all appliednot only psychiatrictreatments. In depression trials, the magnitude of improvement follows the pattern of accepted expectations, when the trial design is known to the investigators, and it is lower in trials with investigators and staff blinded to the trial design and execution. When the level of blinding is high and it is difficult for investigators, raters and patients to guess treatment assignment, the differences between the treatment conditions become small. These assumptions are of paramount importance in interpreting psychedelic studies when blinding is nearly impossible and expectations of therapy are particularly high. For example, findings from a recent trial by Carhart-Harris et al. comparing psilocybin with the standard antidepressant escitalopram were interpreted that psilocybin was at least on a par with the comparator antidepressant. Apart from the fact that the study was not sufficiently powered to exclude a statistical type-2 (beta) error, this interpretation is inconsistent with the existing evidence from studies with antidepressants. At assessment of the primary endpoint, at six weeks, the total HAM-D score had decreased on average by 27.7 percent with escitalopram treatment. This represents a smaller change than that usually observed with placebo in escitalopram trialsor in antidepressant trials in general. Similar observations hold true for the Montgomery-Asberg Depression Rating Scale (MADRS). Thus, in the trial bypatients treated with escitalopram had a symptom improvement that was at most on placebo level and fell short by approximately 50 per cent of the established magnitude of the antidepressant effect of escitalopram as found by. Different from other clinical trials with psychotropic drugs, for which potential patients are approached by investigators, patients usually self-register for trials with psychedelics. In an atmosphere of great public euphoria about these substances, patients associate them with great hopes for an alleviation of their suffering. Patients who wish to be treated with a selective serotonin reuptake inhibitor (SSRI) such as escitalopram do not participate in a clinical trial, but instead consult their psychiatrist or their general practitioner (GP). Due to the unique subjective effects of classical psychedelics at high doses, blinding of treatment with a classical psychedelic is extremely difficult in an RCT. Hence, the most probable reason that patients treated with escitalopram achieved no more than placebo-level improvement was disappointment about assumingly not being in the psilocybin condition. This has likely been evident (or strongly assumed by them) after the first dosing session, before the first escitalopram pill has even been taken. This could be called a "nocebo"-effect, and it underlines the observation that in clinical trials with antidepressants the condition "treatment as usual" sometimes produces worse results than placebo treatment. These considerations have several implications: First, while placebocontrolled studies with a new experimental drug often require a third arm with an active, established comparator in order to demonstrate the "assay sensitivity" of the study, a "real" placebo condition is required in studies of psychedelics with active comparators for the same reason. Second, in order to minimize "nocebo"-effects in comparator arms, treatment with a potentially effective, high dose of the psychedelic should be offered to every patient randomized after assessment of the primary endpoint. This approach of repeated dosing was used in the studies byand, but we propose it as a methodological standard in clinical trials with psychedelics not only for reasons of treatment ethics but also to reduce a potential nocebo effect. In our study, the primary endpoint is assessed immediately before the second dosing session. This proceeding prevents two-thirds of patients from measurement of the primary endpoint being driven solely by disappointment at not having received the supposedly effective, high dose of psilocybin. This also significantly reduces the risk of artificially increasing the dropout rate in the treatment arms with patients who have received the supposedly less or ineffective treatment. In studies that are usually analyzed using the Last-Observation-Carried-Forward (LOCF) method, higher dropout rates in these arms would inflate the effectiveness of the supposedly more effective therapy. This might be especially important in studies of psychedelics for relapse prevention in substance use disorders, in which patients may immediately revert to their harmful use patterns after the experience that they have a high likelihood of not having received the potentially effective therapy. But also in depressed patients, whose condition is inherently characterized by negative thinking, depressed mood, hopelessness and even suicidal ideations or behavior, realizing that one has likely not received the high-dose psychedelic in a one-dose trial design, will most likely lead to a symptom deterioration. Finally, for patients who have often made numerous unsuccessful attempts at treatment and who might have down-titrated their previous antidepressant medication solely for their study participation, it is an ethical imperative to offer them the potentially effective therapy. In our study, every patient is administered at least one high, putatively effective dose of psilocybin. Since one-sixth of patients receive the high dose twice, we also expect an indication of whether two high doses are more effective than one. Attempts have been made to address the problem of blinding, which is very difficult to maintain, in studies with psychedelics by administering so-called "active placebos" or low, potentially ineffective doses of the test substance as comparators. This has also been recommended by the Food and Drug Administration (FDA) and the National Institutes of Health (NIH) during the annual meeting of the American Society of Clinical Psychopharmacology (ASCP; 2019) (www .psychedelic.support/resources/fda-nih-perspectives-psychedelic-drugdevelopment/)). An active placebo -or "concealed placebo" according to its original description in a German article published in 1959-is a placebo that produces subjective side effects that lead the patient to believe that they are receiving active treatment. A compound that has been suggested as an active placebo is niacin. Niacin, or nicotinic acid, a form of vitamin B3, causes peripheral side-effects such as flushing, itching or tingling and abdominal discomfort. Since niacin is not approved in the European Union, we used nicotinamide, the amide of niacin, in our study. Nicotinamide, which can be converted to niacin, causes the niacin-like side-effects usually only at high doses. Another active placebo that has been used to blind the effects of psilocybin in healthy volunteers, is the stimulant methylphenidate. However, healthy subjects and patients usually know the typical effects of psychedelics at least from the literature or the press, and they can relatively easily distinguish them even from psychoactive active placebos (e.g., methylphenidate). An alternative approach that has therefore been proposed is to administer low, supposedly ineffective doses of the experimental compound.administered psilocybin to patients with life-threatening cancer and symptoms of depression and/or anxiety. They compared a high dose (22 or 30 mg/70 kg body weight) with a very low dose (1 or 3 mg/70 kg body weight) and called the latter "placebo-like", still assuming that such low doses are associated with some subjective experience, without having a "therapeutic" effect. The determination of these dosages was based at least in part on a study published by the same authors in 2011.administered five different doses of psilocybin (0, 5, 10, 20, 30 mg/70 kg body weight) to healthy volunteers. They documented significant acute perceptual and subjective effects already at the 5 mg-dose, but only the two highest doses induced "mystical-type" experiences, and only those were associated with long-lasting, persisting effects on attitudes, mood, and behavior. Based on the observations by, we chose the 5 mg-dose of psilocybin as a supposedly therapeutically ineffective dose, while assuming that subjective effects will be noticeable to most patients. Hence, we implemented two comparator arms in order to support the blinding and to test the feasibility of a 5-mg dose as a comparator condition in psilocybin trials. The questions that are touched on here are central to all treatment research with psychedelics. Is a profound psychedelic experience necessary for a long-term therapeutic effect? If so, what dose is necessary and at what dose is the relationship between effectiveness and side effects and risks optimal? How are the therapeutic outcome modulated by the quality of the acute psychedelic experience and what role do the setting and psychotherapeutic interventions play? Which acute and subacute effects of psychedelics are to be considered side effects or AEs and which of themalthough maybe unpleasantare relevant for the therapeutic process and hence beneficial or even required for a treatment response? While dose ranging studies are necessary to determine the minimum effective dose required by regulators, the methodological challenge is substantial and possibly even unsolvable, because there is no other field in pharmacology in which the therapeutic effect seems to be linked as much to the subjective experience of the effect [e.g.. Even innovative substances under development that behave pharmacologically like psychedelic drugs without being associated with a psychedelic subjective effect upon administration to humans will not be able to solve the epistemological problem. If they are therapeutically ineffective, they will not be able to answer the question of whether the psychedelic experience is necessary for the therapeutic effect. On the other hand, if they are effective, they can no longer be called psychedelics. In conclusion, the field of trial methodology faces challenges that are very specific to clinical trials with psychedelic drugs. We are confident that our trial manages to address some of these methodological challenges and hence represents a significant contribution to this rapidly emerging field.